|Publication number||US8116977 B2|
|Application number||US 12/868,626|
|Publication date||Feb 14, 2012|
|Filing date||Aug 25, 2010|
|Priority date||Oct 2, 2008|
|Also published as||CN103080860A, CN103080860B, DE112011102805T5, US20100324772, WO2012027082A1|
|Publication number||12868626, 868626, US 8116977 B2, US 8116977B2, US-B2-8116977, US8116977 B2, US8116977B2|
|Inventors||Gurcan Aral, John W Peake, Brad A Stronger|
|Original Assignee||Trimble Navigation Limited|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (23), Referenced by (4), Classifications (7), Legal Events (2)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This application is a continuation-in-part of U.S. patent application Ser. No. 12/244,198, “Automatic Control of Passive, Towed Implements”, filed on 2 Oct. 2008 and incorporated herein by reference.
The disclosure is related to the control of passive, towed implements for applications such as precision farming.
Farmers in the United States operate over 2 million farms covering roughly one billion acres of land and producing hundreds of billions of dollars of crops each year. The farmers spend tens of billions of dollars per year on seeds, fertilizer, chemicals and fuel. A modern farm is a complex operation where precision and efficiency can have a significant impact on the bottom line. According to the USDA, the most efficient 25% of US corn growers spend about $1 to grow a bushel of corn while growers in the least efficient 25% spend $3 to grow the same amount.
One way farmers improve efficiency is by avoiding unnecessary overlaps in tilling, spraying and harvesting operations. In other words, they avoid driving their tractors and equipment over the same area twice. Consider an 80-acre field and a 44-foot wide sprayer towed behind a tractor as an example. The sprayer is towed across the field in series of overlapping tracks. If the overlap between adjacent sprayer tracks is reduced from two feet to four inches, four acres of spraying are eliminated. Such precision may be achieved by guiding tractors with global positioning system (GPS) based steering systems.
Precision control of passive, towed farm implements such as plows, rippers, disks, planters, applicators, drills and other equipment has other benefits as well. It makes it easier to operate machinery in dark or dusty conditions. Operators can drive faster and reduce driving stress. The quantity of fuel and chemicals used can be decreased, thereby saving money and the environment. Soil compaction can be avoided by keeping heavy equipment on precise tracks.
Advances in GPS technology (and systems based on other global navigational satellite systems (GNSS) such as the Russian GLONASS and the European GALILEO) have made it possible to drive large farm tractors along predetermined paths very accurately. A tractor can return to a field a year after first working it and follow the same track within an inch. The accuracy of a passive, towed implement is not as good, however.
A passive, towed implement does not have its own steering actuators. It is attached to a tractor by a hitch and the tractor pulls it across the ground. The implement may wander off its intended path for any number of reasons including asymmetrical loading (e.g. tougher ground to plow on one side than the other) or drag due to operating on a slope. Skilled tractor operators can compensate for a wandering implement by deliberately steering the tractor away from a desired path so that the implement stays on the path even though the tractor does not. However, despite the best efforts of operators, this manual method is imprecise, takes a long time and travel distance, and causes operator fatigue. It would be better if this “steering away” technique were automated and made more precise.
Modern tractors are often equipped with GPS (or, more generally, GNSS) based autopilot systems. These systems are capable of driving a tractor along a desired path with high accuracy. Further, it is common for such systems to accept offset commands. For example, one may command the system to drive the tractor a specific number of inches left or right of the programmed path. The tractor then travels parallel to, but offset from, the programmed path. Such lateral offsets are called “nudges”.
A wide variety of tractor and autopilot systems are available from different manufacturers. Each of these uses its own control strategy and implementation. The competitiveness of the market ensures that the majority of tractor-autopilot combinations offer path tracking accuracy and response to nudge commands that fall within a relatively narrow range of parameters. Nonetheless, small differences between tractor-autopilot system responses can affect the performance of passive-implement control systems.
What is needed is a control system that ensures that passive, towed implements follow their intended path and correct deviations from the path as quickly as possible. The system should operate in concert with an autopilot-controlled tractor and include a way to measure the tractor-autopilot system response. Such a system would automate the tiring, tedious task of steering a tractor intentionally off path so as to keep an implement on path and improve the accuracy of the actual path followed by the implement.
A control system for passive implements is now described. The system forces a passive, towed implement onto a desired path by directing an autopilot-controlled tractor optimally off the path. The system decreases the response time required for an implement to execute a nudge offset from a predetermined path. The control system calibrates its internal tractor model by measuring the response of the autopilot-controlled tractor to known inputs.
A passive, towed implement's path may wander off course from a desired path for any number of reasons including uneven load on the implement, sloping ground, or random disturbances. One of the tasks of the control system described herein is to minimize implement deviations from a desired path by speeding up the response of an implement to nudge commands.
When the tractor changes heading to move to a new offset, the hitch point initially moves in the opposite direction. This reverse reaction may be modeled by closed form geometric relationships. For example the offset of the hitch point from a desired path is related to the offset of the tractor by: xh=xt−L2 sin ψ where xh is the hitch point offset, xt is the tractor offset, and ψ is the tractor heading.
The GNSS receivers used to measure the position of the tractor and the implement may take advantage of corrections such as those provided by satellite or ground based augmentation systems (SBAS or GBAS). Examples of SBAS include the Federal Aviation Administration's Wide Area Augmentation System (FAA WAAS), the European Geostationary Navigation Overlay Service (EGNOS) operated by the European Space Agency, the Multi-functional Satellite Augmentation System (MSAS) operated by Japan's Ministry of Land, Infrastructure and Transport; and various proprietary systems operated by commercial enterprises. Examples of GBAS include the United States' Local Area Augmentation System (LAAS) and various European differential GPS networks. Even greater accuracy can be achieved by measuring GNSS carrier phase using so-called real time kinematic (RTK) techniques involving a nearby base station located at a surveyed position. RTK allows centimeter-level positioning, for example.
The advanced control system described herein does not depend on detailed knowledge of the tractor autopilot. However, the control system includes the ability to measure the response of the autopilot-controlled tractor to known inputs. Information from response measurements is included in the control system's tractor model.
Given a model for the tractor's dynamics, and having characterized the geometrical parameters of the tractor, hitch and implement, one may design a feedback control model for steering the tractor. For example, tractor motion for small deviations from a desired path may be described by:
where V and L are the tractor's speed and wheelbase respectively and ψ is the tractor's heading. θ is the tractor's steering angle; i.e. the angle of the front wheels away from straight ahead. x is the lateral offset from the desired path. A conventional feedback system may be designed to control a tractor that obeys these equations. When a nudge is introduced to provide a step change in path offset error, the feedback control system changes the tractor's steering angle to make the tractor execute the nudge.
An advanced control system for a passive, towed implement is now described in more detail. The system employs an architecture that is analogous to the Smith predictor introduced in 1957 by O. J. M. Smith. A Smith predictor is most often used in scenarios where a process to be controlled is followed by a delay which prevents immediate measurement of a process value of interest. When a process value of interest is measured after a delay, any actuation to affect the dynamic behavior of the process arrives too late to be used effectively. The Smith predictor circumvents this problem by driving a model of the process and a model of the delay with the same actuation value as applied to the actual process. If the models are accurate, the Smith predictor provides estimated values of otherwise inaccessible immediate process values. It also provides an estimated value of the delayed measurement. The estimated immediate process values are used to modify the dynamic behavior of the process (for example to achieve a faster response) and the estimated delayed process value is compared to the actual delayed measurement. The difference detected in this comparison is used to compensate for process disturbances.
Although the Smith predictor is most often thought of in terms of a process followed by a delay, other dynamic process may be substituted for the delay. In the classic delay scenario, the Smith predictor allows a control system to operate on a process of interest while leaving the delay undisturbed. Here, the Smith predictor is used to control a passive, towed implement while leaving a tractor autopilot system undisturbed.
The model of
Keeping in mind the Smith predictor, one may view the model of
The design of
The design of
Within actual tractor-hitch-implement system 405, “trac”, “hitc”, and “impl” refer to the tractor, hitch and implement respectively. In this actual system nudge commands 420 from control module 460 are directed to the tractor's autopilot system. Movement of the tractor affects movement of the hitch and movement of the towed implement attached to the hitch. Within implement-tractor-hitch model 410, “i
The reordering of elements in the model (i
Within control module 460 error signal 455 is fed to amplifier 461 and multiplier 462. Implement speed 457 is also fed to multiplier 462. The output of the multiplier is fed to discrete-time integrator 464. The outputs of the integrator and of amplifier 461 are summed in adder 463 and fed to amplifier 465. The output of amplifier 465 is nudge signal 420. The gain of amplifier 461 is proportional to the length L1 between the implement hitch point and the implement reference point; i.e. the location of the implement's GNSS receiver. More simply, L1 is the implement boom length. The gain of amplifier 465 is a tuning parameter.
In operation, control module 460 drives error 455 to zero by sending nudge commands to the tractor autopilot within the actual tractor-hitch-implement system 405. Error 455 is composed of immediate implement error 445 (inner loop) and the error (outer loop) or drift between the model 410 of the overall system and the actual system 405. The control module also takes implement speed 457 into account.
There are many ways of characterizing the response of an autopilot-controlled tractor.
Characterizing the response of an autopilot-controlled tractor encompasses sending known, test inputs to the tractor and measuring the resulting tractor motion. The results may then be fit to a model. The model can then be used to predict tractor motion in response to future inputs. The process of constructing models from experimental data is known as system identification and is discussed in detail in, for example, Chapter 12 (“System Identification”) of Digital Control of Dynamic Systems, 3rd edition, by G. F. Franklin, J. David Powell and Michael L. Workman (1998, Addison Wesley Longman, Menlo Park, Calif.), incorporated herein by reference, and Chapter 58 (“System Identification”) of The Control Handbook, W. S. Levine, ed. (1996, CRC Press, Boca Raton, Fla.), incorporated herein by reference.
An example of a pseudorandom input is a long, possibly continuous, stream of very small pulses of random sign and amplitude. Low-level pseudorandom inputs cause tractor motions that are imperceptible to human operators, but that may be extracted from tractor motion data by cross correlation with the known input sequence.
The methods associated with
The advanced control system described here may include application-specific digital electronic circuits or software running in one or more general purpose digital processors. The software executes calculations required to model tractor motion, to evaluate the geometric relationships between the tractor, implement and hitch, and to implement a feedback control system.
In the descriptions above, the desired track has been represented as a series of straight lines; however, the desired track may also contain curves without affecting the design, principles of operation, or efficacy of the control system. In fact, the system improves curve tracking accuracy by directing a tractor to “cut” corners so that as an implement swings wide, it follows a desired path.
The control system for passive, towed implements described herein corrects implement path-tracking errors to zero consistently and optimally. It permits direct precision control of passive implements rather than the tractors that tow them.
The above description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the principles defined herein may be applied to other embodiments without departing from the scope of the disclosure. Thus, the disclosure is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US3141982 *||Jan 6, 1960||Jul 21, 1964||Smith Otto J M||Control system for use in control of loops with dead time|
|US5717593||Sep 1, 1995||Feb 10, 1998||Gvili; Michael E.||Lane guidance system|
|US5928309||Feb 5, 1996||Jul 27, 1999||Korver; Kelvin||Navigation/guidance system for a land-based vehicle|
|US5987383||Apr 28, 1997||Nov 16, 1999||Trimble Navigation||Form line following guidance system|
|US6076612 *||Aug 31, 1999||Jun 20, 2000||Case Corporation||Transition from position to draft mode controlled by hitch position command and feedback|
|US6434462 *||Jun 28, 2001||Aug 13, 2002||Deere & Company||GPS control of a tractor-towed implement|
|US6666279 *||Oct 28, 2002||Dec 23, 2003||New Holland North America, Inc.||Method and apparatus for controlling a tractor/implement combination|
|US6804587 *||Nov 15, 2001||Oct 12, 2004||Integrinautics Corporation||Adjustment of vehicle-implement trajectories to compensate for lateral implement offset|
|US7054731||Aug 29, 2003||May 30, 2006||Trimble Navigation Limited||Farm implement guidance method and apparatus|
|US7065440 *||Jan 22, 2004||Jun 20, 2006||Trimble Navigation, Ltd||Method and apparatus for steering movable object by using control algorithm that takes into account the difference between the nominal and optimum positions of navigation antenna|
|US7142956 *||Jun 24, 2004||Nov 28, 2006||Hemisphere Gps Llc||Automatic steering system and method|
|US7188015||Jul 14, 2004||Mar 6, 2007||Trimble Navigation Limited||Method and system for controlling a mobile machine|
|US7225891 *||May 18, 2004||Jun 5, 2007||Daimlerchrysler Ag||Control system for a vehicle|
|US7263422||Jan 29, 2007||Aug 28, 2007||Trimble Navigation Limited||Method and system for controlling a mobile machine|
|US7373231 *||Jan 8, 2007||May 13, 2008||Hemisphere Gps Llc||Articulated equipment position control system and method|
|US7383114 *||Apr 7, 2006||Jun 3, 2008||Trimble Navigation Limited||Method and apparatus for steering a farm implement to a path|
|US7689354 *||Jul 19, 2005||Mar 30, 2010||Hemisphere Gps Llc||Adaptive guidance system and method|
|US7835832 *||Jan 5, 2007||Nov 16, 2010||Hemisphere Gps Llc||Vehicle control system|
|US7885745 *||Jan 31, 2007||Feb 8, 2011||Hemisphere Gps Llc||GNSS control system and method|
|US20020165669 *||Oct 12, 2001||Nov 7, 2002||Enpoint, L.L.C.||Attitude measurement using a single GPS receiver with two closely-spaced antennas|
|US20060282205 *||Jun 9, 2005||Dec 14, 2006||Lange Arthur F||System for guiding a farm implement between swaths|
|US20080269988 *||Jul 11, 2008||Oct 30, 2008||Feller Walter J||Combined gnss gyroscope control system and method|
|US20090326763 *||Dec 31, 2009||Andrew Karl Wilhelm Rekow||System and method for providing towed implement compensation|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US8655536 *||Nov 16, 2009||Feb 18, 2014||Trimble Navigation Limited||Method and system for augmenting a guidance system with a path sensor|
|US8781685||Jul 17, 2013||Jul 15, 2014||Agjunction Llc||System and method for integrating automatic electrical steering with GNSS guidance|
|US20110118926 *||Nov 16, 2009||May 19, 2011||Trimble Navigation Limited||Method and system for augmenting a guidance system with a path sensor|
|US20110118927 *||Nov 16, 2009||May 19, 2011||Flanders Electric, Ltd.||Systems and methods for controlling positions and orientations of autonomous vehicles|
|U.S. Classification||701/471, 701/50|
|Cooperative Classification||A01B69/004, G05D1/021, G05D2201/0201|
|May 24, 2011||AS||Assignment|
Owner name: TRIMBLE NAVIGATION LIMITED, CALIFORNIA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:ARAL, GURCAN;PEAKE, JOHN W;STRONGER, BRAD A;SIGNING DATES FROM 20110505 TO 20110519;REEL/FRAME:026333/0255
|Jul 29, 2015||FPAY||Fee payment|
Year of fee payment: 4